Evaluation of designs for clinical trials of neuroprotective agents in head injury

In a study involving the statistical modeling of potential head injury trials, we explore approaches to trial design that could enhance their power to detect treatment-related effects on clinical outcome. The study uses a survey organized by the European Brain Injury Consortium of over 1,000 head-injured patients to characterize the population from which trial participants can be selected. A variety of models are postulated for the effects of "neuroprotective" treatment on outcome, and their interaction with a range of strategies for targeting patients for inclusion in the trial is evaluated. A very simple strategy of targeting patients with an intermediate prognosis was found to allow a reduction in sample size by 30%, with no reduction in statistical power. This paper illustrates an important methodology for studying the characteristics of competing trial designs.